IEEE C80216m-08/xxx812457C802.16m-08/xxx
Project / IEEE 802.16 Broadband Wireless Access Working Group <http://ieee802.org/16Title / Advanced downlinkuplink MIMO designSupporting multiple CP sizes in the frame structure
Date Submitted / 2008-057-0577
Source(s) / Hongwei Yang, Hongwei Yang, Xiaolong Zhu, Keying Wu, Liu Hao Liu, Li Dong Li, Xiaolong Zhu,Yang Song, Liyu Cai
Hongwei YANG, Dong LI, Liyu CAI, Keying WU, Xiaolong ZHU
Alcatel Shanghai Bell Co., Ltd. / Voice: + 86-21-50554550; Fax:+ 86-21-50554554
mailto:
Re: / IEEE 802.16m-08/024: Call for Contributions on Project 802.16m System Description Document (SDD)
Topic: “Uplink MIMO”IEEE 802.16m-08/016r1 Call for Contributions on Project 802.16m
System Description Document (SDD)
Target topic: DownlinkUplink MIMO schemes
Abstract / This contribution presents some considerations about downlinkuplink MIMO design for IEEE802.16m, and proposes corresponding text proposalsproposes to support multiple CP sizes in the baseline frame structure so that it can adapt to the environments as well as multicast-broadcast services effectively.
Purpose / To be discussed and adopted by TGm for use in the IEEE802.16m SDD
Notice / This document does not represent the agreed views of the IEEE 802.16 Working Group or any of its subgroups. It represents only the views of the participants listed in the “Source(s)” field above. It is offered as a basis for discussion. It is not binding on the contributor(s), who reserve(s) the right to add, amend or withdraw material contained herein.
Release / The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contribution may be made public by IEEE 802.16.
Patent Policy / The contributor is familiar with the IEEE-SA Patent Policy and Procedures:
http://standards.ieee.org/guides/bylaws/sect6-7.html#6> and <http://standards.ieee.org/guides/opman/sect6.html#6.3>.
Further information is located at <http://standards.ieee.org/board/pat/pat-material.html> and <http://standards.ieee.org/board/pat>.
Advanced downlink MIMO designSupporting multiple CP sizes in the frame Structure
Hongwei Yang, Keying Wu, Liu Hao Liu, Li Dong Li, Xiaolong Zhu, Liyu Cai
Hongwei Yang, Xiaolong Zhu, Keying Wu, Yang Song, Liyu Cai
Hongwei YANG, Dong LI, Liyu CAI, Keying WU, Xiaolong ZHU
Alcatel Shanghai Bell Co., Ltd.
1. 1. Introduction
IEEE802.16m system requires higher performance enhancements of peak data rate, sector throughput and cell edge user throughput, as well as higher mobility and larger coverage. Advanced MIMO techniques will play important roles in satisfying these targets. This contribution presents some basic considerations about downlinkuplink MIMO design.
Overview of MIMO modes
In the IEEE802.16 Session #54 in Orlando, TGm adopted C802.16m-08/118r4 as the 802.16m frame structure baseline content for insertion into SDD update (802.16m-08/003r1). This means that the baseline frame structure supports two different CP sizes of 1/8 and 1/16 Tu, where Tu is the OFDMA symbol duration without CP. However, this frame structure has several open issues:
These two CP sizes are not sufficient for some environments with large delay spread and especially for some services like multicast-broadcast.
The CP size can only change frame by frame rather than sub-frame by sub-frame, which highly restricts the flexibility of system design such as pilot pattern, radio resource scheduling, sub-carrier permutation, HARQ timing, and so on.
The OFDM parameter definition for CP size of 1/16 Tu is not complete. Given the overall OFDM symbols per frame, how to allocate the defined OFDM symbols per sub-frames needs further definition.
This contribution proposes to support another long CP size in the baseline frame structure so that it can adapt to the environments efficiently as well as provide better support for multicast-broadcast services (MBS). Correspondingly, OFDM parameters are defined for long CP, and completed for short CP.
2.
It is not possible to reach all of the challenging targets of 16m system by using a single MIMO mode. Multiple MIMO modes shall coexist in the 16m system to adapt to services and environments. We prefer to design a uniform, multi-mode, adaptive MIMO architecture consisting of single-user MIMO (SU-MIMO), multi-user MIMO (MU-MIMO) and multi-cellBS MIMO as shown in the scenarios of Figure 1:
Figure 1 An application scenario of MIMO modes for 16m system
For SU-MIMO, each user is served by a single BS. It mainly targets for optimizing the single-user-achievable peak data rate.
For MU-MIMO, several users are served by one BS by sharing the same radio resource simultaneously. It is most suitable for heavily loaded systems in cell center, where maximization of overall system throughput is the primary concern.
For multi-cellBS MIMO, multiple users are served by multiple BSs with the same radio resources. In this manner, cell-edge user throughput and cell coverage are improved significantly due to efficient interference reduction through BS coordination.
SU-MIMO
In IEEE 802.16e, SU-MIMO techniques are not defined for uplink. Considering the higher system requirements of IEEE 802.16m and the possibility that some MSs might adopt more than one transmit antenna, SU-MIMO techniques need to be considered for uplink in 802.16m, to improve the service quality experienced by single user.
Some SU-MIMO techniques have been defined in IEEE 802.16e. Considering the higher system requirements on data rate, coverage and mobility, these techniques need further enhancement in the IEEE 802.16m as described in the following subsections. Besides, thesome closed-loop MIMO with the transmitter making useage of the channel information to further enhance the system performance may be considered. Two aspects are critical to closed-loop MIMO, one is the acquisition of (long-term or short-term) channel information at the transmitter, including the quantization and feedback from the terminal; and the other is how to generate the weighting matrix according to the available channel information. The impact of inter-cell interference can also be considered when designing the weighting matrix, for example, using the MIMO plus beamforming technique.
1.1 22Tx
1.2.11 Transmit diversity with 1bit feedback
It is well known in that the sum of two signals can be strengthened if they are in-phase or the absolute phase difference is less than /2. According to this principle, we propose a transmit diversity scheme as shown in Figure 1, which adjusts the phase of the transmitted signal on one channel such that the two branches of signal can be strengthened at the receiver.
Suppose the transmitted signal over Tx-1 is. The transmitted signals from the two antennas are
(1)
where
(2)
with is the channel coefficient from the i-th transmit antenna to the j-th receive antenna.
2.2.11
3.2.11
4.2.11
5.2.11
6.2.11
Figure 1: Uplink transmit diversity scheme with 1-bit feedback
1.1.1 Double-Polarized SFBC
In practical OFDM systems, some frequency offset may exist at the receiver due to some imperfect factors, such as inaccurate frequency synchronization and tracking, large Doppler spread from high mobility etc. The frequency offset will destroy the orthogonality between subcarriers resulting in inter-carrier interference (ICI). Considering the IEEE 802.16m requirement for mobility up to 350 kmph (even 500 kmph in some deployment), ICI may be serious in some application scenarios and significantly degrade the system performance. To solve the ICI problem in cases of large frequency offset, e.g. high mobility scenarios, a double-polarized SFBC (DP-STFC) technique is proposed. The encoding matrix for DP-SFBC can be expressed as follows
(23)
In this expression, the rows correspond to the physically adjacent subcarriers and the columns correspond to the two transmit antennas (here, two transmit antennas are assumed, but the concept can be extended to transmission with more antennas). It can be seen that the DP-SFBC encoding matrix is a staggered arrangement of ordinary SFBC encoding matrix (i.e., positive version) and its negative version. The double-polarized SFBC encoding can be understood as an ICI canceling space-frequency modulation. Correspondingly, at the receiver side, a simple ICI canceling demodulation can be performed firstly, that is, the odd-numbered subcarriers are subtracted from the even-numbered subcarriers for all receive antennas, and then ordinary SFBC detection procedure follows.
2.1 TX
1.2.21 Spatial diversity with rate-1
2. MU-MIMO
In uplink, multiple users can be allocated the same time and frequency resources by BS to transmit their own data traffics, which is called as multi-user MIMO (MU-MIMO). The MU-MIMO transmission is transparent to users. The users sharing the same time-frequency resource can have different numbers and different SU-MIMO modes if they have more than one transmit antennas. Therefore, MU-MIMO has hybrid transmit schemes. Compared with SU-MIMO, MU-MIMO with 2 transmit antennas has an explicit advantage in the sector throughput and cell-edge user throughput. Transmit scheme optimization, multi-user scheduling and UL channel measurement are all performed in BS side.
2.2.21
3. Multi-cell MIMO
3.2.21 In uplink
In legacy system, STBC (space-time block coding) is defined as Matrix A to provide the spatial diversity gain
(Equation 1)
where the ith row gives the symbols transmitted in ith antenna and the jth column gives the symbols transmitted in jth time slot.
Considering the IEEE 802.16m requirement for mobility up to 350 kmph (even 500 kmph in some deployment), STBC is not appropriate for such high speeds due to performance degradation from channel fading in a STBC block. Therefore, the SFBC (space-frequency block coding) should be considered due to robustness to mobility. SFBC also employs the Matrix A format in (Equation 1) but jth column representing the symbols transmitted in jth subcarrier.
STBC is preferred at low mobility where the channel response matrix can be regarded as quasi-static for two consecutive time slots. SFBC is preferred at low delay spread where the channel response matrix is quasi-static for two adjacent subcarriers. To support complicated scenarios in terms of mobility and delay spread, switching between STBC and SFBC can also be considered in the IEEE 802.16m.
3.1.1 Spatial multiplexing with rate-2
In legacy system, SM (spatial multiplexing) is defined as Matrix B to provide the data rate gain
(Equation 2)
where the ith row gives the symbols transmitted in ith antenna.
SM is preferred at high geometry with good channel condition, but it suffers from performance loss under correlated channel due to sensitiveness to channel spatial correlation.
In order to improve the robustness of SM against channel condition as well as to reduce the complexity of the full rate full diversity code, a linearly dispersed SM code can be utilized, such as
(Equation 3)
where . Simulation results show that the new code achieves the same performance as SM for spatially independent channel but much better performance than SM for strongly correlated channel.
3.2 4TX
4.2.21 Spatial diversity with rate-1
CDD (cyclic delay diversity) increases the frequency selectivity of channel by artificial multiple paths, and is especially useful for transferring the control signaling which should be received by all types of terminals. But the performance of CDD is highly dependent on the delay setting. To alleviate the impact of delay setting, antenna permutation can be applied, as shown in the below figure. The mapping pattern between the outputs of CDD and the transmit antennas can vary with time in a certain fashion. For example, in the nth time slot, the ith output of CDD is mapped onto the ith transmit antenna while in the (n+1)th time slot, it is mapped onto the (M-i+1)th antenna, where M=4 is the total number of transmit antennas.
Antenna permutated CDD can also work together with other rate-1 MIMO schemes, such as CDD plus STBC or CDD plus SFBC.
3.2.1 Spatial diversity/multiplexing with rate-2
The combination with antenna permutated CDD with SM shall provide both spatial diversity gain and spatial multiplexing gain with a data rate of 2.
Additionally, with knowledge of the long-term channel information, a layered SM scheme can be considered, which multiplexes several Alamouti-coded matrixes and transmits them simultaneously. Each Alamouti-coded matrix is referred to as a layer. Different layers employ different power levels, which are designed carefully to optimize the system performance and can be implemented with the aid of very limited long-term feedback. This scheme can achieve both the diversity and multiplexing gain. The detection complexity involved is low, which is to the same order as that of the Alamouti code.
The following figure illustrates the transmitter structure of the layered SM for 4Tx. p1 and p2 denote random interleavers while p1 and p2 denote the power levels for layer-1 and layer-2, respectively, and “Mod” the constellation modulator.
3.2.2 Spatial multiplexing with rate-4
The layered SM scheme can be generalized to the rate-4 scenario, by increasing the layer number to four and removing the Alamouti encoder. Different layers still use different power levels. As compared with the Matrix C
(Equation 4)
the layered SM scheme is more robust to the channel spatial correlation, and is more suitable to work with SIC receiver (successive interference cancellation) for performance improvement.
MU-MIMO
With multiple transmit antennas, BS can perform simultaneous transmissions to multiple users over the same radio resource. This technique, known as multi-user MIMO ( MU-MIMO), can be regarded as a generalization of single-user spatial multiplexing where transmissions on multiple spatial channels are dedicated to a single user. Compared with SU-MIMO, MU-MIMO has an advantage in the cell throughput and spectral efficiency, and its performance gain is not limited by the receive antenna number at the terminal side.
MU-MIMO contains three basic technical ingredients: the multi-userMU (multi-user) precoding, multiuseMUr scheduling and multi-userMU channel measurement. Multi-userMU precoding properly weights the data sequences of all users, so that the interference among messages to different users is minimized or effectively controlled. Multi-userU scheduling selects users with the best channel quality and near-orthogonal spatial properties to maximize the overall throughput. MUultiuser channel measurement provides the BS with (part of) the channel information necessary for multiuserMU precoding and scheduling.